# Space Welders β€” Full Content > Complete text of every Space Welders blog post (the 91-episode podcast archive itself is indexed at https://www.spacewelders.com/episodes and syndicated at https://www.spacewelders.com/podcast.xml). Index: https://www.spacewelders.com/llms.txt --- # Hello from the void Published 2026-06-01 β€” https://www.spacewelders.com/blog/hello-space-welders Welcome to the **Space Welders** blog. We're still building, but the lights are on. The short version of why this place exists: we dug 91 episodes of a 2015–2017 tech podcast out of a dead WordPress site, restored the [archive](/episodes), and needed somewhere to write about it. ## What this is A place for notes, build logs, and the occasional deep dive. Expect short, practical posts β€” the kind you'd scribble on the back of a flight plan. ## Markdown works This blog runs on the same markdown engine as the rest of the WiseJNRS sites: - **GFM** tables, task lists, and strikethrough - Fenced code blocks with syntax highlighting - Heading anchors and an auto table of contents ```ts export function ignite(thrust: number) { return `πŸš€ burning at ${thrust}%`; } ``` > Small things, more often. More soon. Strap in. --- # 2015 to 2026: the tech news cycle, then and now Published 2026-06-02 β€” https://www.spacewelders.com/blog/tech-news-then-vs-now > The future has arrived β€” it's just not evenly distributed yet. β€” attributed to William Gibson (1992) Listening back to a tech news show from 2015–2017 is a peculiar experience. You know how every cliffhanger resolves. Some of the era's biggest stories evaporated; others quietly became the world we live in. ## What vanished **The VR land-grab.** For a stretch of 2015–2016, every other headline was a headset. The breathless "this changes everything" energy around Oculus simply isn't there anymore. VR found a niche; it did not eat the living room. **The IoT everything-rush.** "We put a chip in it" was a complete product pitch in 2016. The fridge that tweets, the bin that emails. Most of it is gone, and the warnings the show aired about an IoT bubble look prophetic. **GTA modders vs. Rockstar. Print-the-Legend 3D-printer drama. The "Google Tax."** Genuine front-page stories at the time, now trivia. The half-life of a tech controversy is shorter than anyone admits. ## What stuck **The platform power questions.** Apple's move into streaming, the GST-on-overseas-purchases fight, who controls the pipes: those debates didn't resolve, they escalated. The 2026 version is louder, but it's the same argument. **Encryption and surveillance.** The crypto-wars beat the show covered never ended. It just changed acronyms. **Teaching kids to code.** A recurring, almost romantic theme across the run. Nine years later the question got more complicated, not less, because the thing kids might need to learn keeps moving. ## What changed about the cycle itself The biggest difference isn't the stories, it's the *speed* and the *source*. In 2016 a tech podcast could cover "the week's news" and feel complete. In 2026 the news is continuous, algorithmic, and increasingly written by the same machines it reports on β€” machines that have since [started using the tools themselves](/blog/the-tools-started-using-themselves). Which is the quiet joke of this whole archive: a show about the future, recovered from a forgotten S3 bucket, transcribed by an AI, and republished for an audience that now takes that sentence for granted. The future got here. [Go listen to the run-up.](/episodes) --- ## References - Lopez, L., & Tweel, J. C. (Directors). (2014). *[Print the Legend](https://en.wikipedia.org/wiki/Print_the_Legend)* [Documentary]. Audax Films; Netflix. β€” MakerBot versus Formlabs, plus a cameo from a man printing guns: 2014's most operatic tech feud, now safely behind museum glass. - "[Apple–FBI encryption dispute](https://en.wikipedia.org/wiki/Apple%E2%80%93FBI_encryption_dispute)". *Wikipedia*. β€” The 2016 San Bernardino standoff, filed as an encyclopedia entry because the underlying argument never actually ended. --- # Where to start: a guide to the Space Welders archive Published 2026-06-03 β€” https://www.spacewelders.com/blog/archive-guide > "Begin at the beginning," the King said gravely, "and go on till you come to the end: then stop." β€” Lewis Carroll, *Alice’s Adventures in Wonderland* (1865) The full run is back online: ninety-one episodes of technology, programming, gaming and science, recorded between 2015 and 2017. That's a wall of audio. Here's how to climb it. ## Start at the beginning If you like origin stories, just press play on **Episode 1, "Hello World"** and go in order. The chemistry between Mike and Steve is there from minute one, and you get the news cycle of 2015 unfolding in real time, which is its own kind of time capsule. ## If you want the guests Some of the best episodes bring a third voice into the studio: - **Episode 3, "Comes with Cogs"** β€” with Leon Bambrick (secretgeek), on Star Trek, the "Google Tax", and autonomous cars. - **Episode 10** β€” with Dale Rankine of Future Now Ventures, on startups and the IoT scene. ## If you're here for a theme - **Space:** try **"Australian Astronauts"** and **"Get your ass to Mars"** β€” the episodes where the show fully embraces its name. - **Security & privacy:** **"Impractical Attack Vector"** covers the Cloudflare leak and SHA-1 falling. - **The classics of the era:** anything covering Oculus, E3, or the great "teaching kids to code" debate that ran across several episodes. ## If you just want the vibe Pick a title that makes you laugh and trust it. *"You have to commit to a Chupa-Chup."* *"I don't read anything that doesn't have spank in the title."* *"Cucumbers with anxiety."* The show never took itself too seriously, and the episode names are a reliable mood ring. ## Search it Every episode now has show notes β€” and a growing number have transcripts β€” so the [episodes page](/episodes) is fully searchable. Type a topic, a year, or a guest and follow the thread. Welcome aboard. Strap in. --- ## References - Bambrick, L. (2003–present). *[secretGeek.net](https://www.secretgeek.net/)*. Self-published. β€” Episode 3's guest, still shipping blog posts a decade of internet later, which counts as a longevity record in this medium. - Graham-Cumming, J. (2017, February 23). "[Incident report on memory leak caused by Cloudflare parser bug](https://blog.cloudflare.com/incident-report-on-memory-leak-caused-by-cloudflare-parser-bug/)". *Cloudflare Blog*. β€” The Cloudbleed post-mortem, in which one pointer error sprays other people's cookies across the internet at a rate of 1 in 3.3 million requests. - Stevens, M., Bursztein, E., Karpman, P., Albertini, A., & Markov, Y. (2017). *[SHAttered: The first collision for full SHA-1](https://shattered.io/)*. Google & CWI Amsterdam. β€” Two PDFs, one hash, and a polite academic way of telling the entire industry to stop using SHA-1 immediately. --- # What a 2015 podcast got right (and hilariously wrong) about AI Published 2026-06-04 β€” https://www.spacewelders.com/blog/ai-hits-and-misses > When given sufficient amounts of capacity, training data, and compute time, the representations learned by these models include disentangled features corresponding to high-level concepts. β€” Radford, Jozefowicz & Sutskever (2017) Here's a strange thing to sit with: the show notes for the recovered Space Welders episodes were written by an AI that transcribed the audio and summarised it. The episodes themselves were recorded years before that was remotely possible. The archive is, in a small way, a conversation between two eras of machine learning. So what did Mike and Steve actually see coming? ## The hits **Data science as the main event.** Long before "prompt engineer" was a job title, the show was calling data science "the sexy profession" and digging into how companies were gamifying big-data analysis. The framing was right: the value was moving to whoever could turn data into decisions. **Automation creeping into serious domains.** Automated stock-trading algorithms, recommendation engines, the slow takeover of tasks people assumed were safe. The show treated this as inevitable rather than hypothetical. Correct. **Unsupervised learning as the interesting frontier.** One episode flagged OpenAI's "unsupervised sentiment neuron", a model that learned to track sentiment in text without being told to. At the time it was a neat party trick. In hindsight it was a signpost pointing directly at the large language models that would arrive a few years later. ## The misses **Scale was the whole story, and nobody knew it.** The 2017 assumption was that AI progress would come from cleverer algorithms. The actual answer turned out to be, to a first approximation, *more*: more data, more compute, more parameters. The show can be forgiven; almost everyone missed this. **"AI will help with X" undersold "AI will *do* X."** The era's mental model was AI as an assistant bolted onto existing tools. The reality of 2026 is that AI is often the tool. The leap from augmentation to generation just wasn't on the 2017 radar. **The pace.** Even the optimists assumed a longer runway. The distance from "neat sentiment neuron" to "writes the show notes" was about eight years, not the twenty most people would have guessed. ## The verdict The show understood the *direction* of AI better than most tech media of its day, and badly underestimated the *speed*. Which, honestly, is the most human prediction error there is. Want to hear it for yourself? The episodes where this comes up are scattered through the [archive](/episodes) β€” search "data" or "neuron" and dig in. For the full report card across VR, IoT, crypto and space, see [the predictions scorecard](/blog/we-predicted-the-future). --- ## References - Radford, A., Jozefowicz, R., & Sutskever, I. (2017). "[Learning to Generate Reviews and Discovering Sentiment](https://arxiv.org/abs/1704.01444)" (arXiv:1704.01444). β€” A network reads 82 million Amazon reviews one byte at a time and grows a single neuron that knows how you feel; everything since is arguably a footnote to this party trick. - Davenport, T. H., & Patil, D. J. (2012, October). "[Data Scientist: The Sexiest Job of the 21st Century](https://hbr.org/2012/10/data-scientist-the-sexiest-job-of-the-21st-century)". *Harvard Business Review*. β€” The article that gave a profession its most regrettable epithet, which the profession has spent fourteen years living down. --- # We predicted the future: Space Welders, 2017 vs 2026 Published 2026-06-05 β€” https://www.spacewelders.com/blog/we-predicted-the-future > The best way to predict the future is to invent it. β€” Alan Kay, Xerox PARC (c. 1971) For two years, every week, Mike and Steve sat down to argue about where technology was heading. The show wrapped in 2017. It's now 2026. So: how did they do? We went back through the archive and graded the big calls. ## Virtual reality: **over-hyped** ⭐⭐ No topic dominated the early episodes like VR. Oculus, the consumer Rift, Valve's Lighthouse tracking, Mozilla shipping a VR browser, "AR is the new VR" the week of E3. The show caught the exact peak of the hype. The hindsight: the boom they were watching cooled hard. VR never became the next smartphone. The interesting twist is the one they only half-saw coming, mid-show, in the E3 episode: **AR**, not VR, was the more durable bet, and it took until the 2020s to matter. ## Artificial intelligence: **under-called** ⭐⭐⭐⭐ This is the fascinating one. The show ran *before* the modern AI wave. They talked about "gamifying big data analysis," automated stock-trading algorithms, and an episode literally titled [*The Naked Data Scientist*](/episodes/episode-5) back when data science was "the sexy profession." They were right that machine learning would eat the world. What nobody in 2017 priced in, them included, was how *fast* and how *general* it would get. The OpenAI "unsupervised sentiment neuron" they flagged as a curiosity was an early ancestor of the models that now write show notes for this very archive. (Hi.) We've [scored the AI calls in detail](/blog/ai-hits-and-misses) if you want the full wine-versus-milk breakdown. ## Self-driving cars: **right to be skeptical** ⭐⭐⭐⭐ Google's autonomous cars hitting public roads, the lidar arms race: the show covered the optimism but kept a healthy eyebrow raised. Nine years on, that skepticism aged well. Robotaxis exist in 2026, but the "next year" full autonomy that was perpetually promised in 2017 is still arriving in slices, city by city. ## The Internet of Things: **called the bubble** ⭐⭐⭐⭐⭐ Samsung, Citrix and everyone else piling into IoT; Steve Wozniak warning of an "Internet of Things bubble." The show repeatedly poked at the hype. Verdict: the smart home did arrive, but the indiscriminate "put a chip in everything" gold rush deflated almost exactly as warned. Best call on the board. ## Space: **the optimists won** ⭐⭐⭐⭐ Plenty of rockets, plenty of NASA. The standout is the TRAPPIST-1 episode, seven Earth-size planets, the hosts cheerfully claiming real estate. The space optimism of the show has mostly been vindicated: launch cadence in 2026 would have looked like science fiction in 2017. Mars, of course, is still "a few years out." It always is. ## Crypto: **the blind spot** ⭐⭐ The "Google Tax" and GST on overseas purchases got more airtime than Bitcoin. Understandable in 2015, but it's the one tidal wave the show mostly stood next to without noticing. ## Final grade Strong on skepticism, strongest where they refused to believe the hype (IoT, self-driving). Weakest on the thing that was hardest to see: that a quiet research curiosity would become the defining technology of the decade. Not bad for two guys and a microphone. [Browse the full archive](/episodes) and grade them yourself. --- ## References - Sun, L. (2015, June 10). "[Apple Inc. Co-Founder Steve Wozniak Believes The Internet of Things Bubble Could Burst](https://www.fool.com/investing/general/2015/06/10/apple-inc-co-founder-steve-wozniak-believes-the-in.aspx)". *The Motley Fool*. β€” Woz, on a Sydney stage: "I feel it's kind of like a bubble" β€” the rare tech warning that arrived on time and on the show's home turf. - Radford, A., Jozefowicz, R., & Sutskever, I. (2017). "[Learning to Generate Reviews and Discovering Sentiment](https://arxiv.org/abs/1704.01444)" (arXiv:1704.01444). β€” The "quiet research curiosity" itself: one sentiment neuron, later revealed to be carrying the entire decade. - NASA. (2017, February 22). "[NASA Telescope Reveals Largest Batch of Earth-Size, Habitable-Zone Planets Around Single Star](https://www.nasa.gov/news-release/nasa-telescope-reveals-largest-batch-of-earth-size-habitable-zone-planets-around-single-star/)" (Release 17-015). β€” Seven Earth-size planets around TRAPPIST-1, announced the same week two podcast hosts started claiming real estate 40 light-years out. --- # The tools started using themselves Published 2026-07-05 β€” https://www.spacewelders.com/blog/the-tools-started-using-themselves > Think of MCP like a USB-C port for AI applications. β€” Model Context Protocol documentation Back in 2016, the show ran an episode called [*There's a Bot for that…*](/episodes/episode-36). The joke wrote itself: chatbots were the new apps, every startup had one, and most of them could just about order you a pizza if you typed slowly and forgave often. Scroll the tech feeds this week and the joke has inverted. The trending repositories aren't bots you talk *to*. They're agents that use your tools *for* you. ## What's actually trending A snapshot from this week's feeds, no cherry-picking required: - **Chrome DevTools now ships an MCP server** β€” an official bridge that lets an AI agent open the debugger, inspect the page, and read the network tab the way you would. - **Alibaba's page-agent** drives whole web pages: navigating, clicking, filling β€” the browsing done by the thing that used to be the destination. - **OpenAI publishing plugins for *other* companies' coding agents**, because agent-to-agent tooling is now a category with an audience. - Library releases bragging about **MCP sampling and concurrent tool execution** the way 2016 releases bragged about dark mode. - Think-pieces proposing an **"AI-DLC"** β€” a development lifecycle where the AI isn't a step in the process, it's the process. The common thread is a quiet standard called **MCP** β€” the Model Context Protocol. The lazy analogy is USB for AI tools, and like USB it's boring, which is exactly why it won. Every debugger, browser, wiki and search box gets a port; every agent gets a cable. ## The 2016 mental model, upgraded The bots of the show's era sat *between* you and the tool. You told the bot, the bot told the API, and something vaguely pizza-shaped arrived. The tool was still the tool; the bot was a chatty front door. The 2026 version deletes the middle of that sentence. The agent doesn't relay your intent to the debugger β€” it *is* in the debugger, stepping through frames, reading your console errors, fixing the thing while you get coffee. The front door became the tenant. Mike and Steve's era assumed software's future was more surfaces for humans: more apps, more screens, more dashboards. What actually happened is that the surfaces are increasingly for *machines* β€” and the measure of a good tool in 2026 is how well it can be driven with nobody looking at it. ## The bit the show got right anyway Here's the thing: under the jokes, the bot episode had the right instinct. The interesting question was never "can a bot order pizza." It was "what happens when the interface stops being the product." The 2016 answer was a shrug. The 2026 answer is: the interface becomes an API, the API grows a protocol, and the protocol quietly becomes the most important standard nobody outside the industry has heard of. We [graded the show's big predictions](/blog/we-predicted-the-future) recently and dinged them for missing how *general* AI would get. Fair. But nobody in that scorecard missed this one, because nobody anywhere called it: the tools using themselves wasn't on 2016's bingo card at all. ## The uncomfortable footnote A show recovered from a dead WordPress site, transcribed by a machine, summarised by a machine, and now written about on a site that machines helped build and review. Every layer of this archive has an agent's fingerprints on it β€” [we've noted before](/blog/ai-hits-and-misses) how the show notes themselves are a conversation between two eras of machine learning. At some point the right question stops being "what can the agents do." It becomes "what's left that we insist on doing ourselves." The Welders would have had a good argument about that one. Probably two segments and a tangent about pizza. The bots-for-everything era is in the [archive](/episodes) β€” episode 36 is a good place to hear the future being underestimated in real time. --- ## References - Bynens, M., & Hablich, M. (2025, September 23). "[Chrome DevTools (MCP) for your AI agent](https://developer.chrome.com/blog/chrome-devtools-mcp)". *Chrome for Developers*. β€” Google formally hands the debugger keys to the agents: "AI coding assistants are able to debug web pages directly in Chrome." - Alibaba. (2025). *[page-agent](https://github.com/alibaba/page-agent)* [JavaScript in-page GUI agent, MIT licence]. GitHub. β€” "Control web interfaces with natural language," says the repo description, meaning the browsing is now done by the thing that used to be the destination. - Anthropic et al. (2024–present). *[Model Context Protocol](https://modelcontextprotocol.io/)* [open standard]. β€” The boring little USB-C-for-AI standard (their analogy, officially) underneath everything else on this list; first released November 2024.